In satellite remote sensing in which a growth state and a harvest season of agricultural produce are estimated by sensing reflected light (near-infrared light) from plants using a sensor mounted in a space satellite, it is difficult to acquire data under a night sky or clouds, and it takes several days until data from the satellite comes to hand, and thus, real-time information is hard to get. In addition, since a satellite makes a loop trip and thus getting information of a same spot depends on the cycle of the satellite, rough information of a wide range is obtained, while accurate information of a narrow region is difficult to obtain.
In addition, in near remote sensing that uses a sensor installed on the ground, the distance from a target to the sensor is short, and therefore, there are advantages in that sensing is less affected by the atmosphere than in the satellite remote sensing, data from the target alone can be acquired by the sensor without interference between the sensor and the target, data can be acquired at preferable times, and the like. Such a remote sensing technology in which image information is acquired in proximity to a plant, the image information is transmitted to a computer, a vegetation index is computed by the computer, and a proper harvest time is evaluated or anticipated based on the correlation between the index and evaluation items such as an amount of fiber has been disclosed (refer to PTL 1).